A STUDY OF TRELLIS CODED QUANTIZATION OF SPEECH SIGNALS

Abstract

Current projections for world wide communications in the late 1990's and beyond, point to a proliferation of digital transmi-ssion as a dominant means of communication for voice and data. Digital transmission is expected to provide flexibility, reliability and cost effectiveness with the added potential for communication privacy and security through encryption. The cost of digital storage and transmission media is generally proportional to the amount of digital data that can be stored or transmitted. While the cost of such media decreases every year, the demand for their use Increases at an even higher rate. Therefore there is a continuing need to minimize the number of bits necessary to transmit the signals while maintaining acceptable signal fidelity or quality. The branch of electronics engineering that deals with the later is known as `data compression' or 'coding'. When applied to speech, It is known as speech compression or speech coding [1). Speech coding techniques which can be implemented in real time are of increasing importance today in the medium band bit - rate region of 4-8-16 k bits/second for applications such as mobile radio, satellite channels, source voice terminals for the telephone network, voice/data systems and the emerging ISDN. Many effective coding techniques have been developed which are applicable to code speech signals at such bit rates. In particular,